food_ingredient_parser 1.0.0.pre.9 → 1.0.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: f08e6fcc9422b83503d37b41111f3bd540c11909
-  data.tar.gz: 16af14258ae67fa9b03b2b2196da9631d3dc6a9d
+  metadata.gz: 2c529b63bd3a9f6139ed10663b2cef70ff3d1dc6
+  data.tar.gz: a8effec91559e15920794c61b08a486e572032cf
 SHA512:
-  metadata.gz: b1bfd6c713f0117cc8c13f4110624d270cb471096577e2dc47ea43f6f70664c88c9a2bec444bac407f9424b3c147245e16d04b5b0e9bf8442273c8d90c27955c
-  data.tar.gz: 873de15303ea9bebb4ab9d3504851a4b364e1aef849f6c530235476a4d8e65867fe5426dd7b9274ecd99befb2de83746b8854a6ba08a24cb8caac7525973340b
+  metadata.gz: 35f27c7d83effc16962a65ac4c8c09fb5694373dbd3d2745c434c37ddcf3fc466264c0f10cbe5054876517839a81bf75ab0b1b9876098c1b04b5312138e06ea1
+  data.tar.gz: 28e517777928262b45836d899ff919f725e09ab6e116fd9f58545262b7ae7151821ae683e6d33146d3b5e20f5c02ad7217c4985267460645b0ba80e1ccc19751

data/README.md

@@ -174,6 +174,17 @@ Even though the strict parser would not give a result, the loose parser returns:
 }
 ```
 
+## Compatibility
+
+From the 1.0.0 release, the main interface will be stable. This comprises the two parser's `parse`
+methods (incl. documented options), its `nil` result when parsing failed, and the parsed output's
+`to_h` and `to_html` methods (where available). Please note that parsed node trees may be subject to
+change, even within a major release. Within a minor release, node trees are expected to remain stable.
+
+So if you only use the stable interface (`parse`, `to_h` and `to_html`), you can lock your version
+to e.g. `~> 1.0`. If you depend on more, lock your version against e.g. `~> 1.0.0` and test when you
+upgrade to `1.1`.
+
 ## Test data
 
 [`data/ingredient-samples-nl`](data/ingredient-samples-nl) contains about 150k

data/lib/food_ingredient_parser/loose/scanner.rb

@@ -6,14 +6,14 @@ module FoodIngredientParser::Loose
     SEP_CHARS  = "|;,.".freeze
     MARK_CHARS = "¹²³⁴⁵ᵃᵇᶜᵈᵉᶠᵍªº⁽⁾†‡•°#^*".freeze
     PREFIX_RE  = /\A\s*(ingredients|contains|ingred[iï][eë]nt(en)?(declaratie)?|bevat|dit zit er\s?in|samenstelling|zutaten)\b\s*[:;.]?\s*/i.freeze
-    NOTE_RE    = /\A\b(dit product kan\b|kan sporen\b.*?\bbevatten\b|voor allergenen\b|allergenen\b|E\s*=|gemaakt in\b|geproduceerd in\b|bevat mogelijk\b|kijk voor meer\b|allergie-info|in de fabriek\b|in dit bedrijf\b)/i.freeze
+    NOTE_RE    = /\A\b(dit product kan\b|deze verpakking kan\b|kan sporen\b.*?\bbevatten\b|voor allergenen\b|allergenen\b|E\s*=|gemaakt in\b|geproduceerd in\b|bevat mogelijk\b|kijk voor meer\b|allergie-info|in de fabriek\b|in dit bedrijf\b)/i.freeze
     # Keep in sync with +abbrev+ in the +Common+ grammar, plus relevant ones from the +Amount+ grammar.
-    ABBREV_RE  = Regexp.union(%w[
+    ABBREV_RE  = Regexp.union(/\A(N°|°C|(ijzer|chroom|koper)\s*\(I+\)\s*[[:alnum:]]+)\b/i, *%w[
       a.o.p b.g.a b.o.b c.a c.i d.e d.m.v d.o.c d.o.p d.s e.a e.g e.u f.i.l f.o.s i.a
       i.d i.e i.g.m.e i.g.p i.m.v i.o i.v.m l.s.l n.a n.b n.o n.v.t o.a o.b.v p.d.o
-      p.g.i q.s s.l s.s t.o.v u.h.t v.g v.s w.a w.o w.v vit denat N° °C
+      p.g.i q.s s.l s.s t.o.v u.h.t v.g v.s w.a w.o w.v vit denat
       min max ca
-    ].map {|s| /\A#{Regexp.escape(s)}\b\.?/})
+    ].map {|s| /\A#{Regexp.escape(s)}\b\.?/i}).freeze
 
     def initialize(s, index: 0)
       @s = s                           # input string
@@ -68,7 +68,8 @@ module FoodIngredientParser::Loose
         add_child
       elsif ":".include?(c)       # another open nesting
         if @s[@i+1..-1] =~ /\A\s*(\(|\[)/
-          # ignore if before an open bracket, then it's a regular nesting
+          # ignore colon before an open bracket, then it's a regular nesting
+          name_until_here
         else
           open_parent(auto_close: true)
           @iterator = :colon
@@ -129,16 +130,16 @@ module FoodIngredientParser::Loose
       chars.include?(c) && @s[@i-1..@i+1] !~ /\A\d.\d\z/
     end
 
-    def is_mark?
-      mark_len > 0 && @s[@i..@i+1] !~ /\A°[CF]/
+    def is_mark?(i = @i)
+      mark_len(i) > 0 && @s[i..i+1] !~ /\A°[CF]/
     end
 
-    def mark_len
-      i = @i
-      while @s[i] && MARK_CHARS.include?(@s[i])
-        i += 1
+    def mark_len(i = @i)
+      j = i
+      while @s[j] && MARK_CHARS.include?(@s[j])
+        j += 1
       end
-      i - @i
+      j - i
     end
 
     def abbrev_len
@@ -162,8 +163,8 @@ module FoodIngredientParser::Loose
     end
 
     def add_child
+      name_until_here
       cur.ends(@i-1)
-      cur.name ||= Node.new(@s, cur.interval)
       parent.send(@dest) << cur
       @cur = nil
     end
@@ -192,7 +193,11 @@ module FoodIngredientParser::Loose
     end
 
     def name_until_here
-      cur.name ||= Node.new(@s, cur.interval.first .. @i-1)
+      cur.name ||= begin
+        i, j = cur.interval.first, @i - 1
+        i += mark_len(i) # skip any mark in front
+        Node.new(@s, i .. j) if j > i
+      end
     end
 
     def dot_is_not_sep?

data/lib/food_ingredient_parser/strict/grammar/common.treetop

@@ -44,7 +44,7 @@ module FoodIngredientParser::Strict::Grammar
     end
 
     rule word
-      abbrev / char+
+      abbrev / word_complex / char+
     end
 
     rule and
@@ -104,13 +104,20 @@ module FoodIngredientParser::Strict::Grammar
         'w.a'i /
         'w.o'i /
         'w.v'i /
-        # special words and abbreviations (not auto-generated)
-        'vit.'i /
-        'denat.'i /
-        'N°'i /
-        '°C'i
+        # not auto-generated additions
+        'vit'i /
+        'denat'i
       )
       '.'? ![[:alpha:]]
     end
+
+    rule word_complex
+      # Complex words that contain characters that would otherwise be considered non-words.
+      (
+        'N°'i /
+        '°C'i /
+        ( 'ijzer'i / 'chroom'i / 'koper'i ) ws* '(' 'I'i+ ')' ws* [[:alnum:]]+
+      ) ![[:alpha:]]
+    end
   end
 end

data/lib/food_ingredient_parser/strict/grammar/ingredient_coloned.treetop

@@ -33,7 +33,7 @@ module FoodIngredientParser::Strict::Grammar
 
     # @see Common#word
     rule ingredient_coloned_word
-      abbrev / ( !'/' char )+
+      abbrev / word_complex / ( !'/' char )+
     end
 
   end

data/lib/food_ingredient_parser/version.rb

@@ -1,4 +1,4 @@
 module FoodIngredientParser
-  VERSION      = '1.0.0.pre.9'
-  VERSION_DATE = '2018-09-19'
+  VERSION      = '1.0.0'
+  VERSION_DATE = '2018-09-21'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: food_ingredient_parser
 version: !ruby/object:Gem::Version
-  version: 1.0.0.pre.9
+  version: 1.0.0
 platform: ruby
 authors:
 - wvengen
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2018-09-19 00:00:00.000000000 Z
+date: 2018-09-21 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: treetop
@@ -81,9 +81,9 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
-  - - ">"
+  - - ">="
     - !ruby/object:Gem::Version
-      version: 1.3.1
+      version: '0'
 requirements: []
 rubyforge_project: 
 rubygems_version: 2.6.13